What can you patent in quantum computing?

You can patent a great deal in quantum computing: qubit devices, control and readout electronics, cryogenic hardware, photonic links, fabrication methods, and many control and compilation techniques. The hard cases are pure algorithms. Under U.S. law, a quantum computing patent claim to a mathematical method alone usually fails, while a claim tied to a specific technical improvement in how a machine works often succeeds.

This article is for quantum computing founders and researchers who want to know where that line falls before they file. I cover the eligibility test the USPTO applies today, how it plays out for algorithms, control software and hardware, how to claim hybrid quantum-classical systems, and what to write down early so the application can support the claims you will want later.

Key takeaways

• Hardware inventions (qubit designs, control electronics, cryogenics, interconnects, fabrication) rarely face eligibility problems. Their usual hurdles are novelty, obviousness and enablement.
• Algorithms are judged under the two-step Alice/Mayo framework. A claim to the math alone is at risk; a claim that reflects a concrete improvement to a quantum or classical machine is much stronger.
• Current USPTO guidance, including the precedential Ex parte Desjardins decision, tells examiners to look for technical improvements described in the specification and reflected in the claims.
• Hybrid quantum-classical claims work best when they recite what physically happens: pulses applied, qubits measured, parameters fed back.
• The specification is where eligibility is won. Describe the problem, the mechanism and the measured benefit before you file.

How does patent eligibility work for a quantum computing patent?

Eligibility is a threshold question under 35 U.S.C. § 101. The statute covers any new and useful process, machine, manufacture or composition of matter. Courts add three exceptions: laws of nature, natural phenomena and abstract ideas. Mathematical formulas fall under abstract ideas, which is why quantum algorithms draw scrutiny that a new qubit coupler does not.

The text of Section 101 is short. The exceptions come from the Supreme Court, most importantly Mayo Collaborative Services v. Prometheus Laboratories (2012) and Alice Corp. v. CLS Bank International, decided June 19, 2014.

The two Alice/Mayo steps

In Alice, the Court first asks whether a claim is “directed to” a patent-ineligible concept. If it is, the Court asks what else the claim contains, looking at the elements individually and as an ordered combination, to see whether they transform the claim into a patent-eligible application. The Court also noted that the claims before it did not “purport to improve the functioning of the computer itself.” That remark matters a lot for quantum work.

How the USPTO applies the test today

Examiners follow MPEP § 2106, which breaks the analysis into steps:

1. Step 1: Is the claim to a process, machine, manufacture or composition of matter?
2. Step 2A, Prong One: Does the claim recite a judicial exception, such as a mathematical concept or a mental process?
3. Step 2A, Prong Two: If so, do the other elements integrate the exception into a practical application, for example by improving the functioning of a computer or another technology?
4. Step 2B: If not, do the other elements add “significantly more” than the exception, an inventive concept?

Several recent USPTO documents shape how these steps are applied, and I would read them alongside the MPEP. An August 4, 2025 memorandum to examiners says the mental-process grouping should not be stretched to steps the human mind cannot practically perform, and that a rejection should be made only when ineligibility is more likely than not. A December 5, 2025 advance notice of MPEP changes adopts the reasoning of Ex parte Desjardins, a precedential Appeals Review Panel decision about training a machine learning model. It tells examiners to check whether the specification describes a technical improvement in enough detail and whether the claim includes the elements that deliver it, without evaluating the claim at a high level of generality. The USPTO says these changes will be added to the MPEP in due course.

Applicants can also file evidence. A September 29, 2026 memorandum sets out best practices for optional Subject Matter Eligibility Declarations (SMEDs) under 37 C.F.R. § 1.132. A SMED is a declaration that presents objective evidence aimed only at eligibility, such as why a claimed step improves a technology or cannot practically be performed in the mind.

Algorithms vs. technical improvements: where is the line?

A quantum algorithm claimed as math (a sequence of unitary operations, a cost function, an estimator) is likely to be treated as an abstract idea. The same work is much stronger when the claim reflects how it changes the machine: fewer gates on real hardware, less leakage, faster or more accurate readout, lower classical overhead. The improvement must be real and explained.

The useful question is not “is this software?” but “what technical problem does this solve, and does the claim include the steps that solve it?” Here are examples on each side of the line.

Claims that tend to struggle

• A new variational ansatz claimed as “computing a ground-state energy of a molecule using a parameterized circuit of the following form,” with no link to how it runs better on a device.
• A quantum algorithm for portfolio optimization where the only physical element is “a quantum computer” performing the math. The application is a business problem and the hardware is generic.

Claims that tend to do better

• A compilation method that maps a circuit onto a processor with fixed qubit connectivity and reduces the number of two-qubit gates and SWAP operations, with the reduced gate count and error described in the specification.
• A pulse-shaping method for a transmon that suppresses leakage to non-computational states, claimed as generating and applying specific control waveforms.
• A calibration routine that tracks qubit frequency drift and updates control parameters during operation, reducing downtime between calibrations.
• A readout method that multiplexes several resonators on one line and assigns qubit states with fewer measurement shots.

Each of these recites math somewhere. What helps is that the claim, read as a whole, describes a particular technical solution in a quantum machine rather than the idea of a result.

What quantum hardware can you patent?

Quantum hardware is familiar patent territory. A new device, circuit, apparatus or fabrication process is plainly a machine, manufacture or process, and usually recites no judicial exception at all. The real tests are the ones every physical invention faces: novelty, non-obviousness and enablement of the full claim scope.

Patentable hardware spans every layer of the stack. Common categories include:

• Qubit modalities and devices: superconducting circuits (transmon variants, couplers, resonator geometries), trapped-ion traps and electrode layouts, neutral-atom arrays and optical tweezer setups, spin qubits in silicon or germanium, and photonic qubit sources and detectors.
• Control and readout electronics: microwave and laser control systems, cryogenic CMOS controllers, low-latency feedback on FPGAs, parametric amplifiers, and multiplexed readout chains.
• Cryogenics and packaging: wiring, filtering and thermal anchoring for dilution refrigerators, shielding, and chip packaging that limits crosstalk and loss.
• Photonic interconnects: microwave-to-optical transducers, ion-photon or atom-photon interfaces, and integrated photonic circuits for routing and entanglement distribution. My post on claiming photonic integrated circuits covers that layer in more detail.
• Fabrication: Josephson junction processes, surface treatments that reduce two-level-system loss, and materials or deposition steps that improve coherence.

Novelty, obviousness and enablement

Novelty under § 102 asks whether a single prior-art reference already shows every claim element. Obviousness under 35 U.S.C. § 103 asks whether the claimed invention as a whole would have been obvious to a person of ordinary skill before the effective filing date. In quantum computing, preprints and conference talks are prior art too.

Enablement comes from § 112(a), which requires a written description and enough detail for a skilled person to make and use the invention. In Amgen v. Sanofi (2023), the Supreme Court said “the more one claims, the more one must enable.” A claim to “any qubit with a coherence time above X” invites that problem. I discuss the case in my post on materials patent enablement after Amgen; the same logic applies to qubit materials and device families.

How should you claim hybrid quantum-classical inventions?

Most useful quantum systems today are hybrid. A classical computer prepares parameters, a quantum processor runs circuits, and classical code reads measurements and decides what to do next. Claims to these systems are strongest when they recite the physical interaction between the two sides, not only the math that passes between them.

In practice, I look for ways to tie the claim to concrete events in the hardware:

• Recite control signals applied to identified qubits, measurements taken, and how results change the next control signals.
• Claim the system as well as the method: a quantum processor, control electronics and a classical processor configured to perform the steps.
• Explain the technical benefit in the specification, such as fewer circuit executions, lower latency in a feedback loop, or reduced sensitivity to noise, and keep the claim elements that deliver it.
• Draft fallback claims at several levels of detail: a general method, a version tied to a specific modality, and a version with the specific pulse or circuit structure.

Error mitigation and quantum error correction follow the same principles. Decoders, syndrome-extraction circuits and real-time feedback hardware raise their own questions, which I cover in a separate post on quantum error correction patents.

How do patentability considerations compare across invention types?

Different quantum inventions face different main risks. Hardware usually turns on prior art and enablement. Algorithms usually turn on eligibility first. Control software and hybrid systems sit in the middle, and how the application is written often decides the outcome. The table summarizes typical considerations, not fixed rules.

What should you write down before filing?

The record you create before filing decides much of what you can claim later. Eligibility, enablement and inventorship all depend on what the specification says. A short, technical write-up made while the work is fresh is worth more than a polished summary prepared just before a deadline.

1. The implementation. Describe how it actually runs: the device, pulse sequences, circuit layouts, control flow and parameter values.
2. The problem and the data. State what failed before and show the improvement: fidelities, gate counts, error rates, latency or yield. Data turns an asserted improvement into a described one.
3. Alternatives. List the other modalities, materials, geometries and parameter ranges that also work. These support broader claims and fallbacks.
4. Where the invention lives. Note which parts run in the quantum hardware, the control stack and classical software. This shapes how claims are split.
5. Who contributed what. Record who conceived each claimed feature. Inventorship is a legal question, and it is easier to answer while memories are fresh.
6. Planned disclosures. Note upcoming preprints, talks and investor meetings. My post on when to file before a paper or conference explains why the clock starts earlier than many teams expect.

Investors will ask about these same points during diligence. If a financing is coming, my post on what investors ask about your patents lists the usual questions. You can also read more about how I approach quantum computing patent work.

Frequently asked questions

Can you patent a quantum algorithm?

Sometimes. A claim to an algorithm as pure mathematics is likely to be treated as an abstract idea under 35 U.S.C. § 101. A claim that ties the algorithm to a specific technical improvement in how a quantum or classical machine operates, such as fewer gates on a given processor or lower readout error, has a much better chance. The specification must explain that improvement in technical detail.

Is quantum hardware patentable?

Yes. Qubit devices, control and readout electronics, cryogenic systems, photonic interconnects and fabrication processes are machines, manufactures or processes, and they rarely raise eligibility problems. The usual questions are whether the design is new, whether it would have been obvious from published work, and whether the application teaches a skilled person how to make and use the full range of what is claimed.

Does the Alice decision apply to quantum computing?

Yes. Alice Corp. v. CLS Bank sets the two-step test for every claim that recites an abstract idea, including quantum algorithms and control software. Hardware claims usually do not recite an abstract idea, so the test rarely limits them. For software-heavy claims, the key question is whether the claim reflects an improvement to the functioning of a computer or another technology.

What is a Subject Matter Eligibility Declaration?

It is an optional declaration under 37 C.F.R. § 1.132 that presents evidence on eligibility, for example why a claimed step cannot practically be done in the human mind or why the claims reflect a technical improvement. USPTO memoranda from 2025 and 2026 describe best practices. It supports a good specification but cannot replace one.

Should I patent my quantum software or keep it a trade secret?

It depends on whether others could detect use. Hardware and methods visible in a product or paper are good patent candidates. An internal compiler heuristic or calibration routine that competitors cannot observe may be better kept as a trade secret, since a patent publishes the method. Many quantum companies use both, decided feature by feature.

Talking it through

If you are deciding what to file before a paper, a conference or a fundraise, I am happy to talk it through.

This article is general information, not legal advice, and reading it does not create an attorney-client relationship. Patent law changes and every situation differs. Please talk to a patent attorney about yours.

Message Alexander on LinkedIn